Sequential Feature Selection for Efficient Landslide Segmentation from Multi-Spectral Data

2026-05-10

AI Summary

This study addresses the high redundancy among input channels in multispectral and topographic data, which hampers the efficiency, performance, and interpretability of landslide segmentation models. To tackle this issue, the authors propose an interpretable channel selection framework based on Sequential Floating Forward Selection (SFFS), employing a lightweight U-Net++ proxy model to iteratively construct and prune a feature pool. The method integrates Sentinel-2, ALOS PALSAR, and 16 engineered indices, optimizing channel combinations by explicitly modeling inter-feature interactions rather than evaluating channels in isolation. This approach yields a compact subset of only eight physically interpretable channels that achieves segmentation F1 performance on the Landslide4Sense benchmark comparable to or exceeding that of models using all 30 original channels.

Abstract

Landslide detection from satellite imagery has advanced through deep learning, yet most models rely on large, highly correlated spectral-topographic inputs whose contributions remain poorly understood. The question of which channels are actually necessary has received surprisingly little attention. This matters: redundant or correlated inputs obscure physical interpretability, inflate computational overhead, and can actively degrade model performance through the Hughes Phenomenon. We present a systematic, explainable channel-selection framework for the Landslide4Sense benchmark, combining Sentinel-2 multispectral and ALOS PALSAR terrain data with 16 engineered spectral and structural indices. Rather than relying on conventional single-band drop tests, which evaluate channels in isolation and miss interaction effects, we apply Sequential Forward Floating Selection (SFFS) to iteratively build and prune a candidate feature pool using a lightweight U-Net++ proxy model. Beyond identifying a compact 8-channel subset that matches or exceeds the segmentation F1 of configurations using up to 30 channels, we use the selection process itself to interrogate which spectral and topographic features landslide models genuinely rely on, and what this reveals about the physical cues driving their predictions. We argue that SFFS represents a principled feature selection approach to input design in Earth observation, in contrast to the prevailing practice of appending every available band and hoping the model learns what to ignore.

Problem

• landslide segmentation
• feature selection
• multispectral data
• input redundancy
• model interpretability

Innovation

• Sequential Forward Floating Selection
• Landslide Segmentation
• Feature Selection
• Multispectral Data
• Interpretability